- Email: [email protected]
Genetic Aspects of Non-Polypoid Colorectal Neoplasms
Genetic Aspects of Non-Polypoid C o l o re c t a l N e o p l a s m s Lyn Sue Kahng,
MD
KEYWORDS Colorectal cancer Non-polypoid colorectal neoplasms Research Molecular genetics
Rapid advances in colorectal cancer (CRC) research continue to provide a deeper understanding of the genetic underpinnings of tumor behavior and the ways in which genotype and phenotype may correlate with neoplastic morphology and clinical prognosis. Genetic alterations in the development of colorectal cancer (CRC) are now known to involve different pathways. They initially were characterized in the adenoma–carcinoma sequence, where, in a model first proposed by Fearon and Vogelstein,1 CRC develops as mutations accumulate in a stepwise manner. Characteristic genetic changes include the progressive loss of wild-type tumor suppressor genes; frequently loss of heterozygosity (LOH) at chromosome 5q (APC), 17p (p53), and 18q (DCC/SMAD locus); and activating point mutations of K-ras. Other cancers arise through mutations in the DNA mismatch repair system, as seen in Lynch syndrome, and display microsatellite instability. In addition to these genetic changes, epigenetic modifications have emerged as a crucial factor in CRC development, specifically aberrant promoter hypermethylation, which affects key tumor suppressor genes (CpG island methylator phenotype, or CIMP). The type and degree of genetic instability as well as CIMP status in CRCs allows their classification into different molecular subtypes.2 It is of great interest how pathways of tumor development differ; for example, the genetics of the serrated adenoma pathway are being elucidated and are described separately in this issue. The lower abundance of flat neoplasms has provided more limited and sometimes conflicting data in understanding the genetic factors that give rise to them. Still, genetic differences between non-polypoid and polypoid neoplasms were already proposed as early as 1994 and continue to be explored in the context of newly discovered mechanisms of tumorigenesis for various lesions including flat adenomas, carcinomas, de novo carcinoma, and laterally spreading tumors.
Section of Digestive Diseases and Nutrition, University of Illinois at Chicago and Jesse Brown VA Medical Center, Chicago, IL 60612, USA E-mail address: [email protected] Gastrointest Endoscopy Clin N Am 20 (2010) 573–578 doi:10.1016/j.giec.2010.03.004 1052-5157/10/$ – see front matter. Published by Elsevier Inc.
giendo.theclinics.com
574
Kahng
K-ras MUTATIONS AND THE RAS PATHWAY
The mutations best-characterized in non-polypoid lesions are those in K-ras. Several studies, including the following, have suggested that K-ras mutations are rarer in nonpolypoid lesions than polypoid ones. For example, Yamagata and colleagues found that K-ras mutations could be found in only 23% of flat adenomas, versus 67% of polypoid adenomas. Kaneko and colleagues3 similarly studied a series of 42 carcinomas. They found that p53 overexpression did not differ between the two morphologies but that the non-polypoid cancers lacked K-ras mutations, whereas they were present in 44% of the polypoid cancers. Olschwang and colleagues4 analyzed a series of 44 flat colorectal neoplasms for microsatellite instability and mutations in APC, K-ras, and TGF-RII and found that only K-ras mutations had a lower frequency than in polypoid neoplasms. Umetani and colleagues5 found K-ras mutations in none of their superficial depressed adenomas but 31% of polypoid adenomas. Laterally spreading tumors (LSTs) also have been analyzed, although the results are less clear. For K-ras mutations, Hiraoka and colleagues6 and Mukawa and colleagues7 found a lower prevalence in flat nongranular lesions (16% and 26% respectively) versus granular lesions that are more protruded (78% and 77% respectively). In contrast, Takahashi and colleagues8 found that although K-ras mutation was present in 35% of flat-type LSTs, it was only present in 13% of protruded-type adenomas. In addition to concluding that K-ras mutations correlate with polypoid growth, Yashiro and colleagues9 also found an association of LOH at chromosome 3p with cancers of the de novo type, a region known to contain multiple tumor suppressors or related genes, including MLH1, b-catenin, TGFBR2, and RASSF1A. RASSF1A is a member of a relatively recently discovered family of proteins with tumor suppressor functions, for which epigenetic inactivation by promoter hypermethylation has been described in a wide variety of cancers.10,11 Although almost universal in some cancers such as breast or small cell lung cancer, RASSF1A methylation is less frequent in CRC.10 Van Engeland and colleagues12 found that 45 of 222 (20%) sporadic CRCs had RASSF1A methylation, and of six normal epithelial samples from cancer patients, only one had RASSF1A methylation. Oliveira and colleagues13 found a higher frequency (22 of 51 cases, or 43%), studying polypoid tumors with microsatellite instability. Sakamoto and colleagues14 investigated the frequency in 48 flat tumors comprising 39 early carcinomas and nine high-grade dysplasias; 39 of 48 (81.3%) had RASSF1A methylation, but only 7 of 48 (14.6%) had K-ras mutations. In the 39 cases of tumors with RASSF1A methylation, 19 (49%) also showed RASSF1A methylation in morphologically normal mucosa. It is unclear whether this represents an abnormal background giving rise to tumors, versus a field effect. More data are needed to clarify the roles of these family members, as a separate study by Noda and colleagues15 showed a low (16.4%) incidence of RASSF1A methylation in all tumors, with no difference between flat and polypoid lesions. Analysis of RASSF2 methylation has been found in 43% of tumors in another series, again with no difference between flat and protruded neoplasms.16 Still, alternate pathways to perturbing RAS signaling may be present in flat colonic neoplasms and their background mucosa. ROLE OF APC AND LOH AT CHROMOSOME 17P
There are differing data regarding APC, as Umetani and colleagues5 found that in addition to a lower K-ras frequency, APC mutation was also less frequent in flat versus polypoid adenomas (13% vs 43%, encompassing both depressed and elevated flat adenomas vs polypoid ones) although the frequency was similar in carcinomas.
Genetic Aspects of Non-Polypoid Colorectal Neoplasms
On the other hand, Kaneko and colleagues3 found that although the rate of APC mutation in polypoid versus non-polypoid carcinomas was similar, the types of mutations differed. Non-polypoid carcinomas completely lacked frameshift mutations that were found in 66% of polypoid carcinomas, thus leading the authors to propose that different types of APC mutations could influence tumor morphology and development. In a similar vein, several groups have analyzed loss of heterozygosity at multiple loci, most notably chromosome 17p. Several studies have concluded that LOH at chromosome 17p or p53 overexpression occurs with similar frequency in both flat and polypoid neoplasms.9,17–20 LOH at 17p, however, has been found to be the most frequent (92%) of multiple LOH found in a series of flat tumors,21 and Mueller and colleagues22 found LOH at 17p in 73% of de novo cancers versus 37% of ex-adenoma cancers. OTHER MOLECULAR CORRELATES OF TUMOR BEHAVIOR
With respect to other markers, Wlodarczyk and colleagues23 also found more de novo cancers with decreased E-cadherin expression and extensive stromelysin-3 expression. It has recently been proposed that CD10, b-catenin, and mucin expression may correlate with flat tumor morphology and prognosis also.24 CD10 is a marker that in CRC correlates with a higher incidence of venous invasion or liver metastasis, although a causal relationship has not been established. In contrast, the absence of MUC5AC may be important. MUC5AC is not normally expressed in the colon, but frequently is expressed de novo in adenomas and colorectal cancers. It has been suggested that MUC5AC negativity correlates with higher metastatic potential and poorer prognosis, and it may be differentially expressed in MSI-H (77%) versus MSS cancers (28%).25 Finally, the b-catenin signaling pathway has key roles in development and cancer, where the protein is often stabilized due to APC or b-catenin mutation and aberrant signaling ensues. Koga and colleagues24 performed immunohistochemistry to assess CD10, nuclear b-catenin, MUC2, and MUC5AC in 111 flat colorectal neoplasms and 96 polypoid ones. CD10 was found in 50% of flat low-grade neoplasia (LGN) but 0% of polypoid LGN, as well as 59% of flat high-grade neoplasia versus 33% of polypoid neoplasia. Invasive lesions had, respectively, 51% and 39% CD10 positivity. Nuclear b-catenin was found in 86% of non-polypoid LGN versus 58% of polypoid LGN, but similar percentages of HGN and invasive neoplasia. For nonpolypoid versus polypoid lesions, MUC5AC was present in 25% versus 50% of LGNs; 0% versus 28% of HGN; and 3% versus 32% of invasive neoplasia. Markers of more aggressive phenotype therefore may be present at an earlier stage in flat neoplasms compared with polypoid ones. MICROSATELLITE INSTABILITY AND METHYLATION
Microsatellite instability has been observed by Kaneko and colleagues3 to be similar between polypoid and non-polypoid carcinomas (16% of tumors of each type). In a different study, however, Ogawa and colleagues26 found non-polypoid cancers to have a higher level of MSI with chromosome 17 markers (33%) than polypoid cancers (10%); in the stroma, these numbers were 8% and 4% respectively. Of 16 flat/ depressed lesions described by Kinney and colleagues27 in an American series, five (31%) displayed MSI, similar to the Ogawa study. Thus MSI is not a predominant hallmark of these flat lesions, although further comparisons between non-polypoid and polypoid lesions will show whether there are reproducible differences between the two. Methylation may be important in their development, as Hiraoka and colleagues6
575
576
Kahng
found CIMP-high status less likely in flat LST (8%) than granular LST (61%), and Takahashi and colleagues8 observed a trend toward less gene methylation in flat-type adenomas than protruded-type. FUTURE DIRECTIONS: GENOME-WIDE TRANSCRIPTIONAL ANALYSIS AND DEVELOPMENT OF ANIMAL MODELS
Can animal models and gene expression analyses shed additional light on the development of flat neoplasms? A microarray analysis has been performed on 12 patients with flat adenomas, using adjacent normal mucosa as a control.28 The authors found that 180 genes were differentially expressed, and the expression profiles of right colon lesions were different from those in the left colon. Three genes, MMP7, CDH3, and DUOX2, were up-regulated more than 10-fold in flat adenomas. The authors correlated this transcriptional upregulation with increased immunohistochemical staining of MMP7 and CDH3 in lesions. Real-time polymerase chain reaction analyzing expression of these three genes in polypoid adenomas showed that their levels of expression were elevated over those in normal mucosa, but were still 27% to 58% lower than in flat adenomas. The roles in development of neoplasia are still not fully understood, but it has been suggested that MMP7 overexpression correlates with more advanced adenoma histology, and elevated serum levels with poor prognosis in CRC.29,30 This study is interesting in identifying candidates for further study as well as adding to the literature suggesting differential gene expression in the proximal and distal colon for both normal and neoplastic mucosa.31 One group has reported that in the azoxymethane mouse model of CRC, genetic background can lead to different polyp morphologies; 19% of tumors in KK/HIJ mice are flat, whereas 83% are flat in I/LNJ mice.32 The authors performed serial colonoscopy and histologic analysis, finding that flat polyps continued to grow without becoming more exophytic, and protuberant polyps were overtly raised from first observation. One difference between this study and work described previously was that all azoxymethane-induced flat cancers had adenomatous components, and were thus not de novo cancers. In addition, both flat and polypoid lesions had a similarly low frequency of K-ras mutation (7%), lower than normally found in human polypoid neoplasms. Despite these contrasts, it is an interesting model in which to study the influence of genetics on carcinogenesis and morphology. SUMMARY
Colorectal cancer is now understood to be a heterogeneous disease arising through multiple possible pathways. Elucidating the genetic factors controlling molecular phenotype, morphology, histology, and prognosis of different tumor types continues to be a challenge. Non-polypoid colonic neoplasms provide exciting opportunities for ongoing study of their underlying genetic abnormalities and molecular phenotypes. The varied data from different groups, however, highlight the need for further studies in different populations. With growing awareness of non-polypoid lesions in Western populations, larger series will no doubt be forthcoming to facilitate this research. REFERENCES
1. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990; 61:759–67.
Genetic Aspects of Non-Polypoid Colorectal Neoplasms
2. Jass JR. Classification of colorectal cancer based on correlation of clinical, morphological, and molecular features. Histopathology 2007;50:113–30. 3. Kaneko K, Fujii T, Kato S, et al. Growth patterns and genetic changes of colorectal carcinoma. Jpn J Clin Oncol 1998;28:196–201. 4. Olschwang S, Slezak P, Roze M, et al. Somatically acquired genetic alterations in flat colorectal neoplasias. Int J Cancer 1998;77:366–9. 5. Umetani N, Sasaki S, Masaki T, et al. Involvement of APC and K-ras mutation in nonpolypoid colorectal tumorigenesis. Br J Cancer 2000;82:9–15. 6. Hiraoka S, Kato J, Tatsukawa M, et al. Laterally spreading type of colorectal adenoma exhibits a unique methylation phenotype and K-ras mutations. Gastroenterology 2006;131:379–89. 7. Mukawa K, Fujii S, Takeda J, et al. Analysis of K-ras mutations and expression of cyclooxygenase-2 and gastrin protein in laterally spreading tumors. J Gastroenterol Hepatol 2005;20:1584–90. 8. Takahashi T, Nosho K, Yamamoto H, et al. Flat-type colorectal advanced adenomas (laterally spreading tumors) have different genetic and epigenetic alterations from protruded-type advanced adenomas. Mod Pathol 2007;20: 139–47. 9. Yashiro M, Carethers JM, Laghi L, et al. Genetic pathways in the evolution of morphologically distinct colorectal neoplasms. Cancer Res 2001;61:2676–83. 10. Donninger H, Vos MD, Clark GJ. The RASSF1A tumor suppressor. J Cell Sci 2007; 120:3163–72. 11. Richter AM, Pfeifer GP, Dammann RH. The RASSF proteins in cancer; from epigenetic silencing to functional characterization. Biochim Biophys Acta 2009;1796: 114–28. 12. van Engeland M, Roemen GM, Brink M, et al. K-ras mutations and RASSF1A promoter methylation in colorectal cancer. Oncogene 2002;21:3792–5. 13. Oliveira C, Velho S, Domingo E, et al. Concomitant RASSF1A hypermethylation and KRAS/BRAF mutations occur preferentially in MSI sporadic colorectal cancer. Oncogene 2005;24:7630–4. 14. Sakamoto N, Terai T, Ajioka Y, et al. Frequent hypermethylation of RASSF1A in early flat-type colorectal tumors. Oncogene 2004;23:8900–7. 15. Noda H, Kato Y, Yoshikawa H, et al. Frequent involvement of ras-signaling pathways in both polypoid-type and flat-type early stage colorectal cancers. J Exp Clin Cancer Res 2006;25:235–42. 16. Nosho K, Yamamoto H, Takahashi T, et al. Genetic and epigenetic profiling in early colorectal tumors and prediction of invasive potential in pT1 (early invasive) colorectal cancers. Carcinogenesis 2007;28:1364–70. 17. Hirata I, Wang FY, Murano M, et al. Histopathological and genetic differences between polypoid and nonpolypoid submucosal colorectal carcinoma. World J Gastroenterol 2007;13:2048–52. 18. Rubio CA, Rodensjo M. Mutation of p53 tumor suppressor gene in flat neoplastic lesions of the colorectal mucosa. Dis Colon Rectum 1996;39:143–7. 19. Saitoh Y, Waxman I, West AB, et al. Prevalence and distinctive biologic features of flat colorectal adenomas in a North American population. Gastroenterology 2001; 120:1657–65. 20. Watanabe T, Muto T. Colorectal carcinogenesis based on molecular biology of early colorectal cancer, with special reference to nonpolypoid (superficial) lesions. World J Surg 2000;24:1091–7. 21. Orita H, Sakamoto N, Ajioka Y, et al. Allelic loss analysis of early stage flat-type colorectal tumors. Ann Oncol 2006;17:43–9.
577
578
Kahng
22. Mueller JD, Haegle N, Keller G, et al. Loss of heterozygosity and microsatellite instability in de novo versus ex-adenoma carcinomas of the colorectum. Am J Pathol 1998;153:1977–84. 23. Wlodarczyk J, Bethke B, Mueller E, et al. A comparative study of E-cadherin and stromelysin-3 expression in de novo and ex adenoma carcinoma of the colorectum. Virchows Arch 2001;439:756–61. 24. Koga Y, Yao T, Hirahashi M, et al. Flat adenoma–carcinoma sequence with highmalignancy potential as demonstrated by CD10 and beta-catenin expression: a different pathway from the polypoid adenoma–carcinoma sequence. Histopathology 2008;52:569–77. 25. Byrd JC, Bresalier RS. Mucins and mucin binding proteins in colorectal cancer. Cancer Metastasis Rev 2004;23:77–99. 26. Ogawa T, Yoshida T, Tsuruta T, et al. Genetic instability on chromosome 17 in the epithelium of non-polypoid colorectal carcinomas compared to polypoid lesions. Cancer Sci 2006;97:1335–42. 27. Kinney TP, Merel N, Hart J, et al. Microsatellite analysis of sporadic flat and depressed lesions of the colon. Dig Dis Sci 2005;50:327–30. 28. Kita H, Hikichi Y, Hikami K, et al. Differential gene expression between flat adenoma and normal mucosa in the colon in a microarray analysis. J Gastroenterol 2006;41: 1053–63. 29. Kirimlioglu H, Kirimlioglu V, Yilmaz S, et al. Role of matrix metalloproteinase-7 in colorectal adenomas. Dig Dis Sci 2006;51:2068–72. 30. Maurel J, Nadal C, Garcia-Albeniz X, et al. Serum matrix metalloproteinase 7 levels identifies poor prognosis advanced colorectal cancer patients. Int J Cancer 2007;121:1066–71. 31. Birkenkamp-Demtroder K, Olesen SH, Sorensen FB, et al. Differential gene expression in colon cancer of the caecum versus the sigmoid and rectosigmoid. Gut 2005;54:374–84. 32. Uronis JM, Herfarth HH, Rubinas TC, et al. Flat colorectal cancers are genetically determined and progress to invasion without going through a polypoid stage. Cancer Res 2007;67:11594–600.
MD
KEYWORDS Colorectal cancer Non-polypoid colorectal neoplasms Research Molecular genetics
Rapid advances in colorectal cancer (CRC) research continue to provide a deeper understanding of the genetic underpinnings of tumor behavior and the ways in which genotype and phenotype may correlate with neoplastic morphology and clinical prognosis. Genetic alterations in the development of colorectal cancer (CRC) are now known to involve different pathways. They initially were characterized in the adenoma–carcinoma sequence, where, in a model first proposed by Fearon and Vogelstein,1 CRC develops as mutations accumulate in a stepwise manner. Characteristic genetic changes include the progressive loss of wild-type tumor suppressor genes; frequently loss of heterozygosity (LOH) at chromosome 5q (APC), 17p (p53), and 18q (DCC/SMAD locus); and activating point mutations of K-ras. Other cancers arise through mutations in the DNA mismatch repair system, as seen in Lynch syndrome, and display microsatellite instability. In addition to these genetic changes, epigenetic modifications have emerged as a crucial factor in CRC development, specifically aberrant promoter hypermethylation, which affects key tumor suppressor genes (CpG island methylator phenotype, or CIMP). The type and degree of genetic instability as well as CIMP status in CRCs allows their classification into different molecular subtypes.2 It is of great interest how pathways of tumor development differ; for example, the genetics of the serrated adenoma pathway are being elucidated and are described separately in this issue. The lower abundance of flat neoplasms has provided more limited and sometimes conflicting data in understanding the genetic factors that give rise to them. Still, genetic differences between non-polypoid and polypoid neoplasms were already proposed as early as 1994 and continue to be explored in the context of newly discovered mechanisms of tumorigenesis for various lesions including flat adenomas, carcinomas, de novo carcinoma, and laterally spreading tumors.
Section of Digestive Diseases and Nutrition, University of Illinois at Chicago and Jesse Brown VA Medical Center, Chicago, IL 60612, USA E-mail address: [email protected] Gastrointest Endoscopy Clin N Am 20 (2010) 573–578 doi:10.1016/j.giec.2010.03.004 1052-5157/10/$ – see front matter. Published by Elsevier Inc.
giendo.theclinics.com
574
Kahng
K-ras MUTATIONS AND THE RAS PATHWAY
The mutations best-characterized in non-polypoid lesions are those in K-ras. Several studies, including the following, have suggested that K-ras mutations are rarer in nonpolypoid lesions than polypoid ones. For example, Yamagata and colleagues found that K-ras mutations could be found in only 23% of flat adenomas, versus 67% of polypoid adenomas. Kaneko and colleagues3 similarly studied a series of 42 carcinomas. They found that p53 overexpression did not differ between the two morphologies but that the non-polypoid cancers lacked K-ras mutations, whereas they were present in 44% of the polypoid cancers. Olschwang and colleagues4 analyzed a series of 44 flat colorectal neoplasms for microsatellite instability and mutations in APC, K-ras, and TGF-RII and found that only K-ras mutations had a lower frequency than in polypoid neoplasms. Umetani and colleagues5 found K-ras mutations in none of their superficial depressed adenomas but 31% of polypoid adenomas. Laterally spreading tumors (LSTs) also have been analyzed, although the results are less clear. For K-ras mutations, Hiraoka and colleagues6 and Mukawa and colleagues7 found a lower prevalence in flat nongranular lesions (16% and 26% respectively) versus granular lesions that are more protruded (78% and 77% respectively). In contrast, Takahashi and colleagues8 found that although K-ras mutation was present in 35% of flat-type LSTs, it was only present in 13% of protruded-type adenomas. In addition to concluding that K-ras mutations correlate with polypoid growth, Yashiro and colleagues9 also found an association of LOH at chromosome 3p with cancers of the de novo type, a region known to contain multiple tumor suppressors or related genes, including MLH1, b-catenin, TGFBR2, and RASSF1A. RASSF1A is a member of a relatively recently discovered family of proteins with tumor suppressor functions, for which epigenetic inactivation by promoter hypermethylation has been described in a wide variety of cancers.10,11 Although almost universal in some cancers such as breast or small cell lung cancer, RASSF1A methylation is less frequent in CRC.10 Van Engeland and colleagues12 found that 45 of 222 (20%) sporadic CRCs had RASSF1A methylation, and of six normal epithelial samples from cancer patients, only one had RASSF1A methylation. Oliveira and colleagues13 found a higher frequency (22 of 51 cases, or 43%), studying polypoid tumors with microsatellite instability. Sakamoto and colleagues14 investigated the frequency in 48 flat tumors comprising 39 early carcinomas and nine high-grade dysplasias; 39 of 48 (81.3%) had RASSF1A methylation, but only 7 of 48 (14.6%) had K-ras mutations. In the 39 cases of tumors with RASSF1A methylation, 19 (49%) also showed RASSF1A methylation in morphologically normal mucosa. It is unclear whether this represents an abnormal background giving rise to tumors, versus a field effect. More data are needed to clarify the roles of these family members, as a separate study by Noda and colleagues15 showed a low (16.4%) incidence of RASSF1A methylation in all tumors, with no difference between flat and polypoid lesions. Analysis of RASSF2 methylation has been found in 43% of tumors in another series, again with no difference between flat and protruded neoplasms.16 Still, alternate pathways to perturbing RAS signaling may be present in flat colonic neoplasms and their background mucosa. ROLE OF APC AND LOH AT CHROMOSOME 17P
There are differing data regarding APC, as Umetani and colleagues5 found that in addition to a lower K-ras frequency, APC mutation was also less frequent in flat versus polypoid adenomas (13% vs 43%, encompassing both depressed and elevated flat adenomas vs polypoid ones) although the frequency was similar in carcinomas.
Genetic Aspects of Non-Polypoid Colorectal Neoplasms
On the other hand, Kaneko and colleagues3 found that although the rate of APC mutation in polypoid versus non-polypoid carcinomas was similar, the types of mutations differed. Non-polypoid carcinomas completely lacked frameshift mutations that were found in 66% of polypoid carcinomas, thus leading the authors to propose that different types of APC mutations could influence tumor morphology and development. In a similar vein, several groups have analyzed loss of heterozygosity at multiple loci, most notably chromosome 17p. Several studies have concluded that LOH at chromosome 17p or p53 overexpression occurs with similar frequency in both flat and polypoid neoplasms.9,17–20 LOH at 17p, however, has been found to be the most frequent (92%) of multiple LOH found in a series of flat tumors,21 and Mueller and colleagues22 found LOH at 17p in 73% of de novo cancers versus 37% of ex-adenoma cancers. OTHER MOLECULAR CORRELATES OF TUMOR BEHAVIOR
With respect to other markers, Wlodarczyk and colleagues23 also found more de novo cancers with decreased E-cadherin expression and extensive stromelysin-3 expression. It has recently been proposed that CD10, b-catenin, and mucin expression may correlate with flat tumor morphology and prognosis also.24 CD10 is a marker that in CRC correlates with a higher incidence of venous invasion or liver metastasis, although a causal relationship has not been established. In contrast, the absence of MUC5AC may be important. MUC5AC is not normally expressed in the colon, but frequently is expressed de novo in adenomas and colorectal cancers. It has been suggested that MUC5AC negativity correlates with higher metastatic potential and poorer prognosis, and it may be differentially expressed in MSI-H (77%) versus MSS cancers (28%).25 Finally, the b-catenin signaling pathway has key roles in development and cancer, where the protein is often stabilized due to APC or b-catenin mutation and aberrant signaling ensues. Koga and colleagues24 performed immunohistochemistry to assess CD10, nuclear b-catenin, MUC2, and MUC5AC in 111 flat colorectal neoplasms and 96 polypoid ones. CD10 was found in 50% of flat low-grade neoplasia (LGN) but 0% of polypoid LGN, as well as 59% of flat high-grade neoplasia versus 33% of polypoid neoplasia. Invasive lesions had, respectively, 51% and 39% CD10 positivity. Nuclear b-catenin was found in 86% of non-polypoid LGN versus 58% of polypoid LGN, but similar percentages of HGN and invasive neoplasia. For nonpolypoid versus polypoid lesions, MUC5AC was present in 25% versus 50% of LGNs; 0% versus 28% of HGN; and 3% versus 32% of invasive neoplasia. Markers of more aggressive phenotype therefore may be present at an earlier stage in flat neoplasms compared with polypoid ones. MICROSATELLITE INSTABILITY AND METHYLATION
Microsatellite instability has been observed by Kaneko and colleagues3 to be similar between polypoid and non-polypoid carcinomas (16% of tumors of each type). In a different study, however, Ogawa and colleagues26 found non-polypoid cancers to have a higher level of MSI with chromosome 17 markers (33%) than polypoid cancers (10%); in the stroma, these numbers were 8% and 4% respectively. Of 16 flat/ depressed lesions described by Kinney and colleagues27 in an American series, five (31%) displayed MSI, similar to the Ogawa study. Thus MSI is not a predominant hallmark of these flat lesions, although further comparisons between non-polypoid and polypoid lesions will show whether there are reproducible differences between the two. Methylation may be important in their development, as Hiraoka and colleagues6
575
576
Kahng
found CIMP-high status less likely in flat LST (8%) than granular LST (61%), and Takahashi and colleagues8 observed a trend toward less gene methylation in flat-type adenomas than protruded-type. FUTURE DIRECTIONS: GENOME-WIDE TRANSCRIPTIONAL ANALYSIS AND DEVELOPMENT OF ANIMAL MODELS
Can animal models and gene expression analyses shed additional light on the development of flat neoplasms? A microarray analysis has been performed on 12 patients with flat adenomas, using adjacent normal mucosa as a control.28 The authors found that 180 genes were differentially expressed, and the expression profiles of right colon lesions were different from those in the left colon. Three genes, MMP7, CDH3, and DUOX2, were up-regulated more than 10-fold in flat adenomas. The authors correlated this transcriptional upregulation with increased immunohistochemical staining of MMP7 and CDH3 in lesions. Real-time polymerase chain reaction analyzing expression of these three genes in polypoid adenomas showed that their levels of expression were elevated over those in normal mucosa, but were still 27% to 58% lower than in flat adenomas. The roles in development of neoplasia are still not fully understood, but it has been suggested that MMP7 overexpression correlates with more advanced adenoma histology, and elevated serum levels with poor prognosis in CRC.29,30 This study is interesting in identifying candidates for further study as well as adding to the literature suggesting differential gene expression in the proximal and distal colon for both normal and neoplastic mucosa.31 One group has reported that in the azoxymethane mouse model of CRC, genetic background can lead to different polyp morphologies; 19% of tumors in KK/HIJ mice are flat, whereas 83% are flat in I/LNJ mice.32 The authors performed serial colonoscopy and histologic analysis, finding that flat polyps continued to grow without becoming more exophytic, and protuberant polyps were overtly raised from first observation. One difference between this study and work described previously was that all azoxymethane-induced flat cancers had adenomatous components, and were thus not de novo cancers. In addition, both flat and polypoid lesions had a similarly low frequency of K-ras mutation (7%), lower than normally found in human polypoid neoplasms. Despite these contrasts, it is an interesting model in which to study the influence of genetics on carcinogenesis and morphology. SUMMARY
Colorectal cancer is now understood to be a heterogeneous disease arising through multiple possible pathways. Elucidating the genetic factors controlling molecular phenotype, morphology, histology, and prognosis of different tumor types continues to be a challenge. Non-polypoid colonic neoplasms provide exciting opportunities for ongoing study of their underlying genetic abnormalities and molecular phenotypes. The varied data from different groups, however, highlight the need for further studies in different populations. With growing awareness of non-polypoid lesions in Western populations, larger series will no doubt be forthcoming to facilitate this research. REFERENCES
1. Fearon ER, Vogelstein B. A genetic model for colorectal tumorigenesis. Cell 1990; 61:759–67.
Genetic Aspects of Non-Polypoid Colorectal Neoplasms
2. Jass JR. Classification of colorectal cancer based on correlation of clinical, morphological, and molecular features. Histopathology 2007;50:113–30. 3. Kaneko K, Fujii T, Kato S, et al. Growth patterns and genetic changes of colorectal carcinoma. Jpn J Clin Oncol 1998;28:196–201. 4. Olschwang S, Slezak P, Roze M, et al. Somatically acquired genetic alterations in flat colorectal neoplasias. Int J Cancer 1998;77:366–9. 5. Umetani N, Sasaki S, Masaki T, et al. Involvement of APC and K-ras mutation in nonpolypoid colorectal tumorigenesis. Br J Cancer 2000;82:9–15. 6. Hiraoka S, Kato J, Tatsukawa M, et al. Laterally spreading type of colorectal adenoma exhibits a unique methylation phenotype and K-ras mutations. Gastroenterology 2006;131:379–89. 7. Mukawa K, Fujii S, Takeda J, et al. Analysis of K-ras mutations and expression of cyclooxygenase-2 and gastrin protein in laterally spreading tumors. J Gastroenterol Hepatol 2005;20:1584–90. 8. Takahashi T, Nosho K, Yamamoto H, et al. Flat-type colorectal advanced adenomas (laterally spreading tumors) have different genetic and epigenetic alterations from protruded-type advanced adenomas. Mod Pathol 2007;20: 139–47. 9. Yashiro M, Carethers JM, Laghi L, et al. Genetic pathways in the evolution of morphologically distinct colorectal neoplasms. Cancer Res 2001;61:2676–83. 10. Donninger H, Vos MD, Clark GJ. The RASSF1A tumor suppressor. J Cell Sci 2007; 120:3163–72. 11. Richter AM, Pfeifer GP, Dammann RH. The RASSF proteins in cancer; from epigenetic silencing to functional characterization. Biochim Biophys Acta 2009;1796: 114–28. 12. van Engeland M, Roemen GM, Brink M, et al. K-ras mutations and RASSF1A promoter methylation in colorectal cancer. Oncogene 2002;21:3792–5. 13. Oliveira C, Velho S, Domingo E, et al. Concomitant RASSF1A hypermethylation and KRAS/BRAF mutations occur preferentially in MSI sporadic colorectal cancer. Oncogene 2005;24:7630–4. 14. Sakamoto N, Terai T, Ajioka Y, et al. Frequent hypermethylation of RASSF1A in early flat-type colorectal tumors. Oncogene 2004;23:8900–7. 15. Noda H, Kato Y, Yoshikawa H, et al. Frequent involvement of ras-signaling pathways in both polypoid-type and flat-type early stage colorectal cancers. J Exp Clin Cancer Res 2006;25:235–42. 16. Nosho K, Yamamoto H, Takahashi T, et al. Genetic and epigenetic profiling in early colorectal tumors and prediction of invasive potential in pT1 (early invasive) colorectal cancers. Carcinogenesis 2007;28:1364–70. 17. Hirata I, Wang FY, Murano M, et al. Histopathological and genetic differences between polypoid and nonpolypoid submucosal colorectal carcinoma. World J Gastroenterol 2007;13:2048–52. 18. Rubio CA, Rodensjo M. Mutation of p53 tumor suppressor gene in flat neoplastic lesions of the colorectal mucosa. Dis Colon Rectum 1996;39:143–7. 19. Saitoh Y, Waxman I, West AB, et al. Prevalence and distinctive biologic features of flat colorectal adenomas in a North American population. Gastroenterology 2001; 120:1657–65. 20. Watanabe T, Muto T. Colorectal carcinogenesis based on molecular biology of early colorectal cancer, with special reference to nonpolypoid (superficial) lesions. World J Surg 2000;24:1091–7. 21. Orita H, Sakamoto N, Ajioka Y, et al. Allelic loss analysis of early stage flat-type colorectal tumors. Ann Oncol 2006;17:43–9.
577
578
Kahng
22. Mueller JD, Haegle N, Keller G, et al. Loss of heterozygosity and microsatellite instability in de novo versus ex-adenoma carcinomas of the colorectum. Am J Pathol 1998;153:1977–84. 23. Wlodarczyk J, Bethke B, Mueller E, et al. A comparative study of E-cadherin and stromelysin-3 expression in de novo and ex adenoma carcinoma of the colorectum. Virchows Arch 2001;439:756–61. 24. Koga Y, Yao T, Hirahashi M, et al. Flat adenoma–carcinoma sequence with highmalignancy potential as demonstrated by CD10 and beta-catenin expression: a different pathway from the polypoid adenoma–carcinoma sequence. Histopathology 2008;52:569–77. 25. Byrd JC, Bresalier RS. Mucins and mucin binding proteins in colorectal cancer. Cancer Metastasis Rev 2004;23:77–99. 26. Ogawa T, Yoshida T, Tsuruta T, et al. Genetic instability on chromosome 17 in the epithelium of non-polypoid colorectal carcinomas compared to polypoid lesions. Cancer Sci 2006;97:1335–42. 27. Kinney TP, Merel N, Hart J, et al. Microsatellite analysis of sporadic flat and depressed lesions of the colon. Dig Dis Sci 2005;50:327–30. 28. Kita H, Hikichi Y, Hikami K, et al. Differential gene expression between flat adenoma and normal mucosa in the colon in a microarray analysis. J Gastroenterol 2006;41: 1053–63. 29. Kirimlioglu H, Kirimlioglu V, Yilmaz S, et al. Role of matrix metalloproteinase-7 in colorectal adenomas. Dig Dis Sci 2006;51:2068–72. 30. Maurel J, Nadal C, Garcia-Albeniz X, et al. Serum matrix metalloproteinase 7 levels identifies poor prognosis advanced colorectal cancer patients. Int J Cancer 2007;121:1066–71. 31. Birkenkamp-Demtroder K, Olesen SH, Sorensen FB, et al. Differential gene expression in colon cancer of the caecum versus the sigmoid and rectosigmoid. Gut 2005;54:374–84. 32. Uronis JM, Herfarth HH, Rubinas TC, et al. Flat colorectal cancers are genetically determined and progress to invasion without going through a polypoid stage. Cancer Res 2007;67:11594–600.